real-time customer interactions via sms › sites › default › files › ...10 15 20 25 30 35 40...

1

Real-Time Customer Interactions via SMS

April 2017

We completed a pilot project to determine what can be achieved with users

utilizing Juntos’ iterative process.

2

Using the iterative process, this pilot focused on responding to three user-related questions

and three agent-related questions:

• Question 1: Can we establish effective

communication with users via text

message?

• Question 2: Can we increase menu

activity using text message

communication?

• Question 3: Can we promote

registrations for Tigo Money using text

message communication?

• Question 1: Can we establish two-way

communication with agents via test

message?

• Question 2: Can we learn about

agents’ experiences and realities via

text message?

• Question 3: Can we promote KYC

registration by communicating about it

with agents?

Users Agents

We completed a pilot project to determine what can be achieved with users

utilizing Juntos’ iterative process.

3

Throughout the project, we developed metrics to track our progress in answering these

questions and to document relevant learnings:

• Question 1:

• Response rate

• Enrollment rate

• Question 2:

• Average number of menu

transactions

• Activation rate

• Retention rate

• Question 3:

• Number of KYC registrations

• Number of wallet activations

• Question 1:

• Response rate

• Enrollment rate

• Question 2:

• Learnings shared by the agents

• Question 3:

• Number of KYC registrations

Users Agents

We were able to respond to all questions, develop an optimized strategy,

and obtain valuable learnings about users and agents.

Operations

2,920 agents 54,649 users 121 tests

Results

We achieved up to a 32% response rate

with users and 29% with agents.

The optimized strategy for users achieved

a 21% increase in the average number of

menu transactions.

The optimized strategy for agents

achieved a 12% increase in KYC

registrations.

Learnings

Users are interested in learning more

about Tigo Money.

Agents do respond to text message and

desire a relationship of strong

collaboration with Tigo Money.

It is more effective to first converse about

users’ and agents’ realities, with the goal

of building a relationship, before focusing

on promoting the desire financial behavior.

3

Users

5

We quickly learning how to establish relationships with users via text

message communication.

6

11%15%

19%23%

18%

16% 11%

19%19%

21%20%

27% 29%27%

20%21%

32%

29%

0%

5%

10%

15%

20%

25%

30%

35%

Average Highest rates

Response rate: Averages and highest rates throughout the project

YOUR TIGO: To continue receiving free

messages about the benefits of your

mobile wallet send CONTINUE. To leave,

send LEAVE.

Continue

YOUR TIGO: Thank you for being part of

the Tigo community. We appreciate your

loyalty and congratulate you on your

financial efforts!

Metric definition: The response rate represents the percentage of users that received Juntos

messages and responded back with a text message.

Translated from Spanish

Throughout the project we maintained a high level of quality while

communicating with users.

7

96% 92% 98% 96% 96% 97% 95% 95% 95%

0%

20%

40%

60%

80%

100%

Average

Enrollment Rate: Averages throughout the project

YOUR TIGO: A. I very much like

receiving these message or B. I do not

like receiving these messages. Please

send A or B.

I very much like

receiving these

messages

YOUR TIGO: Thank you for your

response. By By knowing more about you

we can better help you. Have a good day.

Metric definition: The Enrollment Rate represents the percentage of users that received Juntos

messages and did not ask to leave the conversation.


Users shared informationw ith us and participated in developing a

relationship with Tigo Money through Juntos.

8

YOUR TIGO: There are many

free activities that you can do to

have fun, such as going to the

park, exercising, or spending

time with your family.

What nice words

what Encouragement

Thank you

YOUR TIGO: Turn off the TV

before sleeping going to sleep

and disconnect chargers when

they are not in use. Many

chargers use energy when

connected.

YOUR TIGO: To continue

receiving free messages about

the benefits of your mobile wallet

send CONTINUE. To leave, send

LEAVE.

Continue

YOUR TIGO: Thank you for

being part of the Tigo

community. We appreciate your

loyalty and congratulate you on

your financial efforts!

What Tigo offers is

great

As a user, I am grateful

for the concept

YOUR TIGO: Have you used

the Tigo Money menu? We

would love to know what you

thought of the experience.

Share your comments with us

via SMS.

I love using tigo

money. It makes a

lot of things easier

for me

YOUR TIGO: Thank you! We

look forward to knowing more

about your experience.


Users: Optimized Strategy

9

Using the optimized strategy, we achieved notable results in increasing the

number of menu transactions.

10

Treatment month Month after treatment

Juntos Control

+18% more

+21% more

N=614 N=314

Average Menu Txns: Monthly Menu Txns per User (JUNTOS vs. Control)

We observed that the

Juntos treatment

achieves impact not

only in the month of

treatment but also the

month after.

Metric definition: The average number of menu transactions represents the average

of menu transactions completed per user in a given month.

11

7.7%

14.0%

9.0%

5.8%

14.5%

7.5%

Treatment month Month after treatment 2 months aftertreatment

Juntos Control

Juntos =

14.7%

Control = 12.6%

Activation Rate: % of users that completed a menu transaction

N=298

Metric definition: To calculate the activation rate, first a group of users who did not complete a

menu transaction pre-treatment is identified. Then, the % of those users who completed a menu

transaction during the month of treatment, the month after treatment, and two months of treatment

is analyzed.

Over three months

We observed that

before being able to

promote an increase

menu transactions, for

certain users, it is

necessary to focus on

“getting them out of

inactivity”.

Using the optimized strategy, we achieved notable results in increasing menu activation.

12

60.0%

47.0%

39.0%

56.0%

40.0%

34.0%

Treatment month Month after treatment 2 months after treatment

Juntos Control

Retention Rate: % of users who continued to complete a menu transaction

Metric definition: To calculate the retention rate, first a group of users who did complete a menu

transaction pre-treatment is identified. Then, the % of those users who continued to complete

menu transactions during the treatment month, the month after treatment, and tow months after

treatment is analyzed.

Retention is

fundamental in

preventing users from

falling inactive.

N=274

Using the optimized strategy, we achieved notable results in increasing menu activity retention.

Users: Tigo Money KYC

Registrations

13

Most notable results in promoting user KYC registrations

14

7.5%

3.5%3.0%

1.5%


Juntos Control

% KYC Registration: % of users who completed a KYC registration

Metric definition: The % KYC Registration represented the % of users in a group that did not

previously have a KYC registration that did complete a KYC registration the month of treatment

and then the % that completed a KYC registration the month after treatment.

N=67 N=341

These results represent

our strategies that were

most successful in

promoting KYC

registrations. They are not

considered optimized

because, given the size of

the project, they were not

replicated and were not

statistically significant.

Most notable results in promoting wallet activation

15

3532

29

23

0

5

10

15

20

25

30

35

40


Juntos Control

Wallet Activation Rate: Number of new wallets

Metric definition: The wallet activation rate represents the number of wallets that were activated

within a group of users who previously did not have that wallet activated.

N=750 N=753

These results represent

our most successful

strategies in promoting

wallet activation. They are

not considered optimized

because, given the size of

the project, they were not

replicated and were not

statistically significant.

Agents

16

We learned that we can communicate with agents, learn about their reality,

and promote KYC registrations, all via text message

17

Given the fact that the size of the group of agents is much smaller than the size of the

group of users, we focused our iterative process in testing three experiences with the

goal of developing a relationship with the agents and learning about them to be able

to share those learnings with the team. We learned with the agents in August 2016.

Survey Method

Agent Experience

Step by Step

Tests: Three experiencesCommunication

metrics and learnings

Behavioral metrics

and learnings

We learned that we

communicate, learn

about agent realities,

and develop

relationships between

Tigo Money and their

agents.

We observed that we can in

fact promote KYC registration

with agents.

We identified that Step by

Step achieves the highest

behavioral metrics while the

other experiences’ content

achieves higher

communication metrics.

We learned how to communicate with agents.

18

20%

13%

19%

21%

14%

29%

0%

5%

10%

15%

20%

25%

30%

35%

August September October

Average Highest rates

Without previous

experience, we were

able to achieve

response rates with

agents that were

almost as high as our

response rates with

users.

Metric definition: The response rate represents the percentage of agents who received Juntos

messages and responded with a text message.

Response rate: Averages and highest rates throughout the project

19

Metric definition: The enrollment rate represents the percentage of users who received Juntos

messages and did not ask to leave the conversation.

Enrollment rate: Averages throughout the project

90%

87%

93%

84%

85%

86%

87%

88%

89%

90%

91%

92%

93%

94%

August September October

Average

Average

YOUR TIGO: What would help your

business most? A. Better attention from

Tigo or B. A better signal to do

transactions. Send A or B.

A


responses. By knowing your needs we

can better help you. We appreciate your

loyalty as a Tigo agent. Have a good

week!

We learned how to communicate with agents.


We developed a communication strategy that achieved results in promoting

KYC registration.

20

Average # KYC Registrations: Average per Agent

Increase of 12%

Our hypothesis is that

agents can be a more

effective channel to

promote KYC

registration (via SMS)

than by going through

the users themselves.

Metric definition: The average number of KYC registrations represents how many KYC

registrations each agent completes on average within the designated gorup for analysis.

N=337 N=357

19.5

20

20.5

21

21.5

22

22.5

23

23.5

24

24.5

Pre-treatment Treatment month

Pre-treatment Treatment month

We learned what topics Tigo Money agents are most interested in covering.

21

Topics of interest

We tested different communication strategies to learn the best way to communicate with agents.

Content about

their business

Different from users, agents less variety in terms of

needs. Agents hope to have a collaborative relationship

with Tigo Money. They were interested in content about

how Tigo could support their business, especially in

growing their business.

Opportunity to

express

themselves

Agents value the opportunity to express their opinion

about Tigo Money anonymously.

Why KYC

They were interested in understanding why

completing KYC registrations with users is important

before wanting to discuss how to carry the

registration process out.

YOUR TIGO: What happens

more in your business? 1. Too

many people withdraw 2. Few

people deposit. Send 1 or 2.


response. Please share with us

how you think Tigo can solve this

problem. Thanks!

2

I am missing more

publicity signs!!!


Agents sharing information with us and participated in developing a

relationship with Tigo Money through Juntos

22

YOUR TIGO: What happens more

in your business? 1. Too many

people withdraw 2. Few people

deposit. Send 1 or 2.

1


response. Please share with us

how you think Tigo can solve this

problem. Thanks!

Having ATMs

To deposit cash

Increasing the

credit with Banco

Sudameris

YOUR TIGO: What type of

business do you have? Tell us

about it. If you need a

business idea send 5, we will

send you a list of ideas.

I need credit 80

million to buy land

YOUR TIGO: ID

registration established a

link between a user’s

mobile wallet and his

personal information,

such as date of birth and

ID number.

It is always necessary

to do. The registration

system does not work

very well

if I have a package it

does not work

you guys need to

facilitate things.

You guys should

Advancing financial inclusion to improve the lives of the poor www.cgap.org

real-time customer interactions via sms › sites › default › files › ...10 15 20 25 30 35 40...

Documents